基于多进程的电力系统频域特征值并行搜索算法  

A Multi-process Based Frequency Domain Eigenvalue Parallel Search Algorithm for Power System Small Signal Stability Assessment

作　　者：仲悟之[1] 宋新立[1] 汤涌[1] 卜广全[1] 吴国旸[1] 刘涛[1] 

机构地区：[1]中国电力科学研究院,北京市100192

出　　处：《电力系统自动化》2010年第21期11-16,共6页Automation of Electric Power Systems

基　　金：国家科技支撑计划重大项目(2008BAA13B08)~~

摘　　要：对大规模电力系统进行小干扰稳定分析时,特别是在线动态安全预警系统中,由于系统线性化矩阵的规模庞大,求解特征值的串行算法难以满足计算速度的要求。近年来,多核CPU技术发展迅速,充分利用多核CPU的计算机硬件资源,采用并行计算技术是大幅提高小干扰稳定计算速度的有效途径。提出了基于多进程的电力系统频域特征值并行搜索算法,并应用于国内广泛使用的电力系统小干扰稳定性分析程序(PSD-SSAP)中。该算法具有计算准确、实现简单、并行计算效率高等优点。实际大规模电力系统的算例测试和分析验证了该算法的正确性和有效性。Small signal stability analysis of a large-scale power grid（in particular on-line dynamic security early warning system） involves high rank differential-algebra equations （DAEs）in system modeling.In such circumstances,serial eigenvalue computation algorithm is computationally expensive and may not meet the requirements of computation speed.In recent years, the development of multi-core CPU technology is very rapid.Effective utilization of the hardware resources of multi-core CPU computers provides an effective way to speed up eigenvalue computation for small signal stability analysis using parallel computing technology.This paper proposes the frequency domain eigenvalue parallel search algorithm of power system small signal stability assessment.This algorithm is based on multi-process which has been realized in the PSD-SSAP widely used in China.The calculation is accurate,the realization of the algorithm is simple,and the parallel computation efficiency is very high.Numerical simulation results of an actual large-scale power systems small signal analysis show that the proposed algorithm is correct and efficient.

关 键 词：并行计算 电力系统 小干扰稳定 多进程 多核处理器 

分 类 号：TM711[电气工程—电力系统及自动化]